Upload QformT5ForVQA

config.json

@@ -1,23 +1,33 @@
 {
   "QformConfig": {
-    "encoder_hidden_size": 2,
-    "hidden_size": 2,
-    "model_type": "blip_2_qformer",
-    "num_attention_heads": 2,
-    "num_hidden_layers": 2
+    "classifier_dropout": null,
+    "model_type": "blip_2_qformer"
   },
   "T5ModelConfig": {
-    "model_type": "t5"
+    "architectures": [
+      "T5ForConditionalGeneration"
+    ],
+    "d_ff": 3072,
+    "d_model": 768,
+    "decoder_start_token_id": 0,
+    "model_type": "t5",
+    "n_positions": 512,
+    "num_decoder_layers": 12,
+    "num_heads": 12,
+    "num_layers": 12,
+    "output_past": true,
+    "torch_dtype": "float32",
+    "vocab_size": 36096
   },
   "architectures": [
     "QformT5ForVQA"
   ],
-  "from_pretrained_qform": null,
-  "from_pretrained_t5": "t5-small",
+  "continue_training": true,
+  "from_pretrained_qform": "Gregor/mblip-mt0-xl",
+  "from_pretrained_t5": "VietAI/vit5-base",
   "model_type": "qformt5",
-  "patch_per_side": 4,
+  "patch_per_side": 14,
   "torch_dtype": "float32",
-  "train_from_original": true,
-  "transformers_version": "4.41.2",
-  "visual_dmodel": 8
+  "transformers_version": "4.36.0",
+  "visual_dmodel": 768
 }

generation_config.json

@@ -1,4 +1,4 @@
 {
   "_from_model_config": true,
-  "transformers_version": "4.41.2"
+  "transformers_version": "4.36.0"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:838a2b1a67003d292ff59c40361f22452e60172969a2fcc864d1b2436b16e4be
-size 242182328
+oid sha256:a42077edf34283c0b46daf44bca32c142fa7f24ba775edf0cf5099c14d8f74d3
+size 1331508408